Search results

The OOM killer will become active and kill a process, usually a VM as they a good targets with their high RAM usage ;) Though you might have other mechanisms that can free up RAM to some degree. ZFD for example might use quite a bit of RAM for caching but will free it when needed.

You do need to setup a reverse proxy on one machine that will then forward requests to the actual machines. This is usually done depending on the domain. So you can setup multiple (sub) domains and all point to the reverse proxy. There are plenty of guides on how to set up reverse proxies. It...

It does exactly that. Replicate the VMs disk to the other node. You can then enable it to be HA so that if the one node fails, it will be booted on the other. Each node in a cluster has the configurations of all guests so the failover node can take ownership of the guest as long as it has...

dd'ing from /dev/zero with compression on shows how fast your CPU can deliver zeros, compress them and how much metadata around it actually needs to be written to disk ;) If you want to do benchmarking checkout fio which will do benchmarking well and give you better insight into IOPS and...

First, dd ist Not a good tool for benchmarks. Second,using ZFS on a RAID controller is not good practice. ZFS wants the disks as naked as possible and a RAID controller abstracts too much from the physical disks and has its own cache which can interfere. Third, ZFS can be configured in many...

Have you checked if there are firmware updates available for the SSDs? Otherwise, if you can spare one SSD, I would perform some benchmark and write tests on it to see if similar behavior can be observed.

While not perfectly correlated time wise this is still interesting: Apr 10 13:06:33 pve-us smartd[4841]: Device: /dev/sda [SAT], SMART Prefailure Attribute: 194 Temperature_Celsius changed from 99 to 97 Apr 10 13:06:33 pve-us smartd[4841]: Device: /dev/sdb [SAT], SMART Prefailure Attribute...

Nothing in the /var/log/syslog or /var/log/kern.log? How are the drives sda-sdd connected? All on the same controller?

Anything in the sys or kernel logs around the times before you see the drop?

@toddimus while your argument is not a 100% technically correct, it is true in this situation. # ipcalc 10.102.0.0/16 Address: 10.102.0.0 00001010.01100110. 00000000.00000000 Netmask: 255.255.0.0 = 16 11111111.11111111. 00000000.00000000 Wildcard: 0.0.255.255...

Spricht etwas gegen den Proxmox VE installer? Wenn nicht wird es wohl am einfachsten sein, du installierst damit nochmal neu und wählst während der Installation (bei der Dateisystemauswahl) ZFS aus mit RAID1 um den Mirror zu bekommen. Dann sparst du dir dieses herumgetue mit MD RAID? auf ein...

user: root pw: whatever you set during installation make sure that the Realm is set to Linux PAM!

Btw, which EPYC model do you have installed? I am just reading up on it again so this is from memory and probably not technically 100% correct. EPYC Naples (1 at the end of the 4 digit product code) has Numa nodes because they had to group the chiplets and not every chiplet has the same fast...

That would have been my next guess and something that I will check on the next boot of that server, if I don't forget until then ;)

which command did you use to get that output? On a AMD EPYC 7302P 16-Core Processor I get the following: # dmesg | grep -i numa [ 0.006392] No NUMA configuration found

Too bad but I guess you need to use the machine in production and cannot tinker around for too long. It would have been interesting to know what the cause was. I doubt that it was the same problem that LTT encountered in that video because you are far from having that many really fast SSDs...

Well, you have it in you container config that it will use vmbr0 as bridge to connect to any other network. Your network config needs to be adapted then. An example config that you would also get if you use the installer or GUI to set up your network would be: iface eno1 inet manual auto...

yes yep having the IP set on the bridge port and not the bridge is not a default setup. that might cause some issue but I am not sure. The webui should still be reachable. After a regular install with the installer that would be the setup you have out of the box. Not sue if it would be the...

AFAIR the number of max backups is set for each storage. Do you back up all VMs to the same storage or to different ones which might have a different max backup setting? The first is the running kernel, the second one the version of PVE packages. You might want to install updates. Those...

As another note, bridges do not need to have an IP set. So if the NIC which is configured as bridge port should only be used for the VMs and not to access the Proxmox UI, just don't configure an IP address. You can also create "internal" networks by not setting a bridge port.